CN104143334B - Programmable graphics processor and its method that audio mixing is carried out to MCVF multichannel voice frequency - Google Patents
Programmable graphics processor and its method that audio mixing is carried out to MCVF multichannel voice frequency Download PDFInfo
- Publication number
- CN104143334B CN104143334B CN201310170251.XA CN201310170251A CN104143334B CN 104143334 B CN104143334 B CN 104143334B CN 201310170251 A CN201310170251 A CN 201310170251A CN 104143334 B CN104143334 B CN 104143334B
- Authority
- CN
- China
- Prior art keywords
- audio
- voice data
- graphics processor
- programmable graphics
- road
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Stereophonic System (AREA)
- Processing Or Creating Images (AREA)
Abstract
A kind of method that audio mixing is carried out the invention discloses programmable graphics processor and its to MCVF multichannel voice frequency, is related to digital processing field.The method includes：The N roads voice data to be played in a period of time is stored to input texture buffer, the N roads voice data of input is bound into the data texturing of programmable graphics processor；Programmable graphics processor is pre-processed to each road voice data simultaneously, at least one operation in preprocessing process according to per the coded system of voice data, sample rate or volume all the way, being decoded accordingly to the road voice data, in resampling or decay；Pretreated each road voice data is superposed to audio signal all the way by programmable graphics processor, the road audio signal is stored to output texture buffer, and exported to playback equipment.The present invention significantly reduces audio mixing time delay using the powerful concurrent operation ability of programmable graphics processor.
Description
Technical field
The present invention relates to digital processing field, more particularly to a kind of parallel processing of utilization programmable graphics processor
Ability carries out the method and programmable graphics processor of audio mixing to MCVF multichannel voice frequency.
Background technology
Audio mixing refers to be input into reproduction of multiple audio streams signal to audio mixing equipment simultaneously, and reproduction of multiple audio streams signal is finally mixed into
The process that audio stream signal is exported all the way.Mixer is generally divided into hardware audio mixing and software audio mixing.
Hardware audio mixing use special sound processing apparatus, such as specialty audio mixing equipment or sound card pre-processed and
Audio mixing.Professional audio mixing equipment, such as Pioneer SVM-1000VDJ mixing desks, it is bulky, it is expensive, it is adaptable to recording studio
Or the occasion such as studio hall, be not suitable for portable personal playback equipment and use.Sound card audio mixing is PC（Abbreviation PC）Most often adopt
Audio mixing mode, its digital audio processing IC chip for passing through customization carries out decoding, resampling, the decay of MCVF multichannel voice frequency,
More advanced sound card can also support that echo suppresses for microphone input part, and ambient noise suppresses, the function such as speech enhan-cement.
But the object of sound card audio mixing is mainly multiple physics such as circuit input, PC loudspeaker, microphone, CD playback, the MIDI playback in PC
The audio that equipment is produced, the audio mixing of the MCVF multichannel voice frequency played simultaneously for multiple audio softwares, then need to be carried out by operating system
Software audio mixing.
Software audio mixing is, using software programming mode, to use central processing unit（CPU）Carry out sound signal processing and audio mixing.
During audio mixing is carried out, audio can also experience the treatment such as decay, resampling.Operating system audio mixing belongs to the one of software audio mixing
Plant implementation, the audio play-back application interface that audio playing program is provided by call operation system（API）To operation
System submit to need play voice data, and by operating system unification carry out audio mixing after, output to sound card or audio broadcasting set
It is standby.Software audio mixing is the stereo process that MCVF multichannel voice frequency is realized using the computing capability of CPU, is not rely on specific hardware.
In the case that the calculated load of operating system is heavier, or concurrent audio way is excessive, such as multiplayer, MPTY etc. will
There is the situation that audio mixing postpones, so as to influence audio mixing effect.
Current many software and hardware manufacturers it is also proposed many above-mentioned mix process of acceleration, reduces audio mixing and acoustic processing postpones
Technology.For example, part has hardware-accelerated sound mixing function using the high-grade independent sound card with audio processing chip, can be obvious
Reduce CPU usage.But this mode can substantially increase the manufacturing cost and volume of equipment, for current PC, intelligence
Mobile phone etc. is generally using integrated on cheap Audio or piece（SoC）The equipment such as sound card are not applied to simultaneously.What Microsoft proposed
DirectSoundTMTechnology, is by providing one group of special API, the audio buffer of direct read/write driver of sound card, so as to reach
To memory access latencies are reduced, the purpose of audio mixing efficiency is improved.But the calculating energy of CPU is still limited by these technological essences
Power.
Therefore, it is necessary to propose a kind of suitable portable equipment, be independent of operating system computing capability and can just reduce audio mixing
The audio mixing technology of time delay.
The content of the invention
An embodiment of the present invention technical problem to be solved is：Propose a kind of utilization programmable graphics processor and
The method that row disposal ability carries out audio mixing to MCVF multichannel voice frequency, is carried out with solving current portable equipment dependence operating system computing capability
Audio mixing, so as to cause asking for audio mixing delay occur in the case of operating system calculated load is heavier or concurrent audio way is excessive
Topic.
The one side of the embodiment of the present invention is provided one kind MCVF multichannel voice frequency is mixed based on programmable graphics processor
The method of sound, including：The N roads voice data to be played in a period of time is stored to input texture buffer, the N roads that will be input into
Voice data binds the data texturing of programmable graphics processor；Programmable graphics processor enters to each road voice data simultaneously
Row pretreatment, according to per the coded system of voice data, sample rate or volume all the way in preprocessing process, to the road audio number
At least one operation according to being decoded accordingly, in resampling or decay；Programmable graphics processor will be pretreated each
Road voice data is superposed to audio signal all the way, and the road audio signal is stored to output texture buffer, and exports to broadcasting
Equipment.
The programmable graphics that audio mixing is carried out to MCVF multichannel voice frequency another aspect provides a kind for the treatment of of the embodiment of the present invention
Device, including：Data texturing binding module, delays for the N roads voice data to be played in a period of time to be stored to input texture
Area is rushed, the N roads voice data of input is bound into the data texturing of programmable graphics processor；Pretreatment module, for simultaneously
Each road voice data is pre-processed, in preprocessing process according to per the coded system of voice data all the way, sample rate or
Volume, the road voice data is decoded accordingly, resampling or decay at least one operation；Superposition output module,
For pretreated each road voice data to be superposed into audio signal all the way, the road audio signal is stored to output texture and is delayed
Area is rushed, and is exported to playback equipment.
The present invention by the way that the N roads voice data of input to be bound the data texturing of programmable graphics processor, using can
The programming powerful concurrent operation ability of graphic process unit, at the same each road voice data is decoded, resampling or decay etc. it is pre-
Treatment, is finally superposed to audio signal all the way by pretreated each road voice data, and output is played out to playback equipment,
Whole mix process eliminates the reliance on operating system computing capability using the powerful concurrent operation ability of programmable graphics processor,
Therefore, even if in the case where operating system calculated load is heavier, programmable graphics processor can also in real time complete audio mixing,
So as to significantly reduce audio mixing time delay, Consumer's Experience is improved.Also, programmable graphics processor can be decoded to audio,
The complex process such as resampling, decay, so as to also improve audio quality on the basis of audio mixing real-time is improved.The present invention is proposed
Audio mixing scheme be particularly suitable for being currently equipped with powerful programmable graphics processor but relatively weak portable many of integrated sound card
Apparatus for media playing, such as PC, smart mobile phone.
By referring to the drawings to the detailed description of exemplary embodiment of the invention, further feature of the invention and its
Advantage will be made apparent from.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
The accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, without having to pay creative labor, may be used also
Other accompanying drawings are obtained with according to these accompanying drawings.
Fig. 1 is the principle schematic that the present invention carries out audio mixing based on programmable graphics processor to MCVF multichannel voice frequency.
Fig. 2 is the stream of method one embodiment that the present invention carries out audio mixing based on programmable graphics processor to MCVF multichannel voice frequency
Journey schematic diagram.
Fig. 3 is that programmable graphics processor of the present invention shows the flow that each road voice data pre-process one embodiment
It is intended to.
Fig. 4 is the structural representation of programmable graphics processor one embodiment that the present invention carries out audio mixing to MCVF multichannel voice frequency
Figure.
Fig. 5 is the structural representation that the present invention carries out another embodiment of the programmable graphics processor of audio mixing to MCVF multichannel voice frequency
Figure.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than whole embodiments.Below
Description only actually at least one exemplary embodiment is illustrative, and never conduct is to the present invention and its application or makes
Any limitation.Based on the embodiment in the present invention, those of ordinary skill in the art are not making creative work premise
Lower obtained every other embodiment, belongs to the scope of protection of the invention.
Fig. 1 is based on programmable graphics processor for the present invention（Abbreviation GPU）The principle that audio mixing is carried out to MCVF multichannel voice frequency is illustrated
Figure.As shown in figure 1, audio mixing refers to be input into reproduction of multiple audio streams signal to audio mixing equipment simultaneously, and reproduction of multiple audio streams signal is final
It is mixed into the process of audio stream signal output all the way.
It is stronger equipped with performance for current most of portable multimedia playback equipment such as PC, smart mobile phone etc.
Big programmable graphics processor（For example, tall and handsome Geforce, Tegra series for reaching, the Radeon series of ATI, the Mali of ARM
Series, the Andreno series of high pass, the gc series of Vivante, PowerVR series of Imagenation etc. support CUDA,
The graphic chips of the GPU general-purpose computations programming languages such as GLSL, OpenCL）, but the relatively weak situation of integrated sound card performance, this hair
It is bright to propose a kind of scheme that Multi-channel audio sound mixing is carried out using programmable graphics processor, it is advantageous that by programmable graphics
The powerful concurrent operation ability of processor, breaks away from the dependence to operating system computing capability, heavier in operating system calculated load
In the case of, programmable graphics processor can also in real time complete audio mixing, so as to audio mixing time delay is greatly reduced, improve user's body
Test.On the other hand, programmable graphics processor can also carry out such as decoding, resampling to audio, decay more complicated treatment,
So as to also improve audio quality on the basis of audio mixing real-time is improved.Audio mixing scheme proposed by the present invention is carried out in detail below
Describe in detail bright.
Fig. 2 is the stream of method one embodiment that the present invention carries out audio mixing based on programmable graphics processor to MCVF multichannel voice frequency
Journey schematic diagram.As shown in Fig. 2 the sound mixing method of the embodiment is comprised the following steps：
Step 102, the N roads voice data to be played in a period of time is stored to input texture buffer, the N that will be input into
Road voice data binds the data texturing of programmable graphics processor.
Assuming that the i-th road voice data（May be by a rules or μ rule compressions）It is Si(j), j=1...ki, then will be per length all the way
It is 1 × kiVoice data be stored as 2 d texture array B（Size is N × K）The 1st of the i-th row to kthiIndividual element, its
In, N represents audio way, and K represents the maximum length of all voice datas.I-th tunnel is input into the corresponding line of n-th sample of audio
Managing coordinate is（I, n）, n-th output corresponding texture coordinate of audio sample point be（1, n）.
Step 104, programmable graphics processor is pre-processed to each road voice data simultaneously, the root in preprocessing process
According to per the coded system of voice data, sample rate or volume all the way, being decoded accordingly to the road voice data, resampling or
At least one operation in decay.
Pretreated each road voice data is superposed to audio signal all the way by step 106, programmable graphics processor, will
The road audio signal is stored to output texture buffer, and is exported to playback equipment.
According to the parallel behavior of programmable graphics processor, the calculating of each output audio sample point is executed in parallel
's.Therefore, a computational methods for output sample are exemplarily only given herein, and the computational methods of other output samples can join
Examine the computational methods that this exports sample.
Fig. 3 is that programmable graphics processor of the present invention shows the flow that each road voice data pre-process one embodiment
It is intended to.As shown in figure 3, the preprocessing process of the embodiment is comprised the following steps：
Step 202, before being pre-processed to each road voice data, needed for n-th output audio sample point of calculating
Coordinate of the sample value of the i-th tunnel input audio in texture buffer is input into（X, y）,
X=i,
y=n×ki/ N,
Wherein, kiThe number of elements of audio is input into for the i-th tunnel, i ∈ N, x are integer, and y is integer or decimal, and N represents audio
Way.
Step 204, according to coordinate（X, y）Carry out re-sampling operations.
Due to managing the effect of buffering area similar to array, can be accessed according to the coordinate specified（Read/write）Wherein deposit
Numerical value.But difference is, if the coordinate for reading is not integer, programmable graphics processor can be according to neighbouring rounded coordinate position
The sample value interpolation put obtains the value of the position.
A kind of exemplary re-sampling operations proposed by the present invention include：
（1）Obtained from input texture buffer and be located at coordinateWithSample value at 2 points is respectively s1
And s2, wherein,Expression is rounded downwards,Expression rounds up.
A kind of texture mapping method is, it is assumed that the i-th road voice data（May be by a rules or μ rule compressions）It is Si(j), j=
1...ki, then will be 1 × k per length all the wayiVoice data be stored as 2 d texture array B（Size is N × K）I-th row
1st to kthiIndividual element, wherein, N represents audio way, and K represents the maximum length of all voice datas.I-th tunnel is input into sound
The corresponding texture coordinate of n-th sample of frequency is（I, n）, n-th output corresponding texture coordinate of audio sample point be（1, n）.
（2）If s1And s2Without coding, according to formulaInterpolation arithmetic is carried out to obtain
Arrive（X, y）Sample value s；
（3）If s1And s2It is encoded, according to formulaInterpolation arithmetic is carried out to obtain
（X, y）Sample value s, s'1It is s1Decoded sample value, s'2It is s2Decoded sample value.
Step 206, if s1And s2It is encoded, before resampling, decoding operate is carried out to voice data.
A kind of exemplary decoding operate proposed by the present invention includes：
（1）A rules or/and μ rule decodings look-up table array are bound the data texturing of programmable graphics processor, the texture
Positioned at coordinate（1, u）The value at place is the corresponding decoding outputs of u, and the span of u is [0-255].A rule decoding look-up table such as tables 1
Shown, table 1 invests specification last page.Similar, according to the corresponding relation of sample value before and after μ rates coding, μ rates can be constructed
Decoding look-up table.
（2）Calculate audio sample values s1In the coordinate of a rules or μ rule decoding look-up table array correspondences texture buffer（u1,
v1）, u1=1, v1=s1, according to coordinate（u1, v1）Restrained by a or μ rule decoding look-up tables decode and obtain decoded sample value
s'1。
（3）Calculate audio sample values s2In the coordinate of a rules or μ rule decoding look-up table array correspondences texture buffer（u2,
v2）, u2=1, v2=s2, according to coordinate（u2, v2）Restrained by a or μ rule decoding look-up tables decode and obtain decoded sample value
s'2。
Step 208 is right（X, y）Sample value s decayed, specifically include：
Will（X, y）Sample value s be multiplied by the attenuation factor of the i-th tunnel audioiAudio sample values after being decayed, and will
Audio sample values after decay are superimposed to output sample o.The formula realized using computer program is expressed as：o=o+αi×s。
Sound mixing method of the invention is further elaborated with reference to instantiation.
Assuming that carry out audio mixing to two-way audio, a length of 10ms during the audio of buffering area buffering, the sample rate of the first via is
44kHz, attenuation coefficient is 0.4, remembers that its buffering area array is S1(j), j=1...44.The sample rate on the second tunnel is 48kHz, decay
Coefficient is 0.6, remembers that its buffering area array is S2(j), j=1...48.
（1）The corresponding array of two-way audio is stored in one 2 × 48 two-dimentional buffering area array, and binds programmable figure
Shape processor data texturing, is designated as arr2d_audio_tracks.A rule decoding look-up tables are stored in the buffering area of 1 × 256
In array, and programmable graphics processor data texturing is bound, be designated as arr1d_a_law_lookup_table.
（2）I=1 is made, the 1st tunnel audio is processed
a）Calculate current output sample point corresponding coordinate (x, y) in audio texture is input into.Current output sample point
Coordinate is determined by GPU, it is possible to use mPosition.x, mPosition.y conduct interviews.Be calculated according to below equation (x,
y)。
y=mPosition.y*arr1d_audio_len[i]/out_len
=44/48*mPosition.y
x=mPosition.x
Y1=floor（y）// y is rounded downwards
Y2=ceil（y）// y is rounded up
b）The sample value s1, s2 of (x, y1) and (x, y2) place are obtained by texture mapping to arr2d_audio_tracks：
s1=texture2D(arr2d_audio_tracks,vec2(x,y1));
s2=texture2D(arr2d_audio_tracks,vec2(x,y2));
Texture2D may be referred to abovementioned steps 204 and obtain s1, the method for s2 herein.
c）The type of coding of first via audio is judged according to arr1d_audio_enc_type [i], finds its value
It is A_LAW, uses a rule codings.Look-up table then is restrained to s1 using a, and s2 is decoded, i.e., to arr1d_a_law_
It is the value at (1, s1) place that lookup_table obtains coordinate by texture mapping, and again to s1 assignment.I.e.：
s1=texture2D(arr1d_a_law_lookup_table,vec2(1,s1))
s2=texture2D(arr1d_a_law_lookup_table,vec2(1,s2))
The method that texture2D may be referred to the decoding of abovementioned steps 206 herein.
d）Resampling is carried out to first via audio, the interpolation method for using is linear interpolation.Then first via audio is inserted
The mathematical formulae of value result is described as：
e）Attenuation coefficient arr1d_audio_attenuation_coeff [i] according to the i-th tunnel audio decays to s,
And it is superimposed to output sample value mFragColor.
mFragColor=mFragColor+arr1d_audio_attenuation_coeff[i]*s
（3）I=2 is made, the 2nd tunnel audio is processed, repeat the above steps a）-e）.
For above-mentioned example, a kind of pretreatment implementation process of use GLSL language description is given here：
uniform sampler arr2d_audio_tracks;// two-dimensional array, is input into audio, and often row represents sound all the way
Frequently
uniform sampler arr1d_a_law_lookup_table;// one-dimension array, a rule decoding look-up tables
uniform int out_len;// output audio sample points
uniform int arr1d_audio_len[MAX];// one-dimension array, record is per road audio length
uniform int arr1d_audio_enc_type[MAX];// one-dimension array, coding staff of the record per road audio
Formula
uniform float arr1d_audio_attenuation_coeff[MAX];// one-dimension array, record is per road
The attenuation coefficient of audio
void main(){
for(i=0;i<MAX;i++){
float x,y,y1,y2;
float s,s1,s2;
// calculate current output sample point corresponding coordinate (x, y) in audio texture is input into
y=mPosition.y*arr1d_audio_len[i]/out_len;
x=mPosition.x;
Y1=floor（y）;// y is rounded downwards
Y2=ceil（y）;// y is rounded up
s1=texture2D(arr2d_audio_tracks,vec2(x,y1));
s2=texture2D(arr2d_audio_tracks,vec2(x,y2));
if(arr1d_audio_enc_type[i]==A_LAW){
// perform a rule decodings
s1=texture2D(arr1d_a_law_lookup_table,vec2(1,s1))
s2=texture2D(arr1d_a_law_lookup_table,vec2(1,s2))
}
// perform resampling
s=(y-y1)*s2+(y2-y)*s1;
// perform decay and be superimposed
mFragColor=mFragColor+arr1d_audio_attenuation_coeff[floor(x)]*s;
}
Above-mentioned sound mixing method, by the way that the N roads voice data of input to be bound the data texturing of programmable graphics processor,
Using the powerful concurrent operation ability of programmable graphics processor, while being decoded to each road voice data, resampling or being declined
The pretreatment such as subtract, pretreated each road voice data is finally superposed to audio signal all the way, and output is entered to playback equipment
Row is played, and whole mix process eliminates the reliance on operating system meter using the powerful concurrent operation ability of programmable graphics processor
Calculation ability, therefore, even if in the case where operating system calculated load is heavier, programmable graphics processor can also be complete in real time
Into audio mixing, so as to significantly reduce audio mixing time delay, Consumer's Experience is improved.Also, programmable graphics processor can enter to audio
The complex process such as row decoding, resampling, decay, so as to also improve audio quality on the basis of audio mixing real-time is improved.This
The audio mixing scheme that invention is proposed is particularly suitable for being currently equipped with powerful programmable graphics processor but integrated sound card is relatively weak
Portable multimedia playback equipment, such as PC, smart mobile phone.
Fig. 4 is the structural representation of programmable graphics processor one embodiment that the present invention carries out audio mixing to MCVF multichannel voice frequency
Figure.As shown in figure 4, the programmable graphics processor of the embodiment includes：
Data texturing binding module 302, for the N roads voice data to be played in a period of time to be stored to input texture
Buffering area, the data texturing of programmable graphics processor is bound by the N roads voice data of input；
Pretreatment module 304, for being pre-processed to each road voice data simultaneously, according to each in preprocessing process
The coded system of road voice data, sample rate or volume, are decoded, in resampling or decay accordingly to the road voice data
At least one operation；
Superposition output module 306, for pretreated each road voice data to be superposed into audio signal all the way, by the road
Audio signal is stored to output texture buffer, and is exported to playback equipment.
Fig. 5 is the structural representation that the present invention carries out another embodiment of the programmable graphics processor of audio mixing to MCVF multichannel voice frequency
Figure.
As shown in figure 5, programmable graphics processor also includes：Coordinate calculation module 408, for entering to each road voice data
Before row pretreatment, the sample value of the i-th tunnel input audio needed for calculating n-th output audio sample point is in input texture buffering
Coordinate in area（X, y）,
X=i, y=n × ki/ N,
Wherein, kiThe number of elements of audio is input into for the i-th tunnel, i ∈ N, x are integer, and y is integer or decimal.
As shown in figure 5, pretreatment module 304 includes resampling unit 3041, for being obtained from input texture buffer
Positioned at coordinateWithSample value at 2 points is respectively s1And s2,Expression is rounded downwards,Represent upward
Round；If s1And s2Without coding, according to formulaInterpolation arithmetic is carried out to obtain（X,
y）Sample value s；If s1And s2It is encoded, according to formulaInterpolation arithmetic is carried out to obtain
Arrive（X, y）Sample value s, s'1It is s1Decoded sample value, s'2It is s2Decoded sample value.
As shown in figure 5, pretreatment module 304 includes decoding unit 3042, if for s1And s2It is encoded, carrying out
Before interpolation arithmetic, a rules or/and μ rule decodings look-up table array are bound the data texturing of programmable graphics processor；Calculate
Audio sample values s1In the coordinate of a rules or μ rule decoding look-up table array correspondences texture buffer（u1, v1）, u1=1, v1=s1, according to
Coordinate（u1, v1）Restrained by a or μ rule decoding look-up tables decode and obtain decoded sample value s'1；Calculate audio sample values
s2In the coordinate of a rules or μ rule decoding look-up table array correspondences texture buffer（u2, v2）, u2=1, v2=s2, according to coordinate（u2, v2）
Restrained by a or μ rule decoding look-up tables decode and obtain decoded sample value s'2。
As shown in figure 5, pretreatment module 304 includes attenuation units 3043, for inciting somebody to action（X, y）Sample value s be multiplied by the i-th tunnel
The attenuation factor of audioiAudio sample values after being decayed, and the audio sample values after decay are superimposed to output sample.
The above-mentioned programmable graphics processor for carrying out audio mixing, programmable figure is bound by by the N roads voice data of input
The data texturing of shape processor, using the powerful concurrent operation ability of programmable graphics processor, while to each road voice data
Decoded, the pretreatment such as resampling or decay, pretreated each road voice data is finally superposed to audio signal all the way,
And output is played out to playback equipment, whole mix process utilizes the powerful concurrent operation ability of programmable graphics processor,
Operating system computing capability is eliminated the reliance on, therefore, even if in the case where operating system calculated load is heavier, at programmable graphics
Reason device can also in real time complete audio mixing, so as to significantly reduce audio mixing time delay, improve Consumer's Experience.Also, programmable graphics
The complex process such as processor can be decoded to audio, resampling, decay, so as to be gone back on the basis of audio mixing real-time is improved
Improve audio quality.Audio mixing scheme proposed by the present invention be particularly suitable for being currently equipped with powerful programmable graphics processor but
The relatively weak portable multimedia playback equipment of integrated sound card, such as PC, smart mobile phone.
One of ordinary skill in the art will appreciate that realizing that all or part of step of above-described embodiment can be by hardware
To complete, it is also possible to instruct the hardware of correlation to complete by program, described program can be stored in a kind of computer-readable
In storage medium, storage medium mentioned above can be read-only storage, disk or CD etc..
The foregoing is only presently preferred embodiments of the present invention, be not intended to limit the invention, it is all it is of the invention spirit and
Within principle, any modification, equivalent substitution and improvements made etc. should be included within the scope of the present invention.
Table 1-a rule decoding look-up tables
| A rule inputs | Decoding output | A rule inputs | Decoding output | A rule inputs | Decoding output | A rule inputs | Decoding output |
| 0 | -5504 | 64 | -344 | 128 | 5504 | 192 | 344 |
| 1 | -5248 | 65 | -328 | 129 | 5248 | 193 | 328 |
| 2 | -6016 | 66 | -376 | 130 | 6016 | 194 | 376 |
| 3 | -5760 | 67 | -360 | 131 | 5760 | 195 | 360 |
| 4 | -4480 | 68 | -280 | 132 | 4480 | 196 | 280 |
| 5 | -4224 | 69 | -264 | 133 | 4224 | 197 | 264 |
| 6 | -4992 | 70 | -312 | 134 | 4992 | 198 | 312 |
| 7 | -4736 | 71 | -296 | 135 | 4736 | 199 | 296 |
| 8 | -7552 | 72 | -472 | 136 | 7552 | 200 | 472 |
| 9 | -7296 | 73 | -456 | 137 | 7296 | 201 | 456 |
| 10 | -8064 | 74 | -504 | 138 | 8064 | 202 | 504 |
| 11 | -7808 | 75 | -488 | 139 | 7808 | 203 | 488 |
| 12 | -6528 | 76 | -408 | 140 | 6528 | 204 | 408 |
| 13 | -6272 | 77 | -392 | 141 | 6272 | 205 | 392 |
| 14 | -7040 | 78 | -440 | 142 | 7040 | 206 | 440 |
| 15 | -6784 | 79 | -424 | 143 | 6784 | 207 | 424 |
| 16 | -2752 | 80 | -88 | 144 | 2752 | 208 | 88 |
| 17 | -2624 | 81 | -72 | 145 | 2624 | 209 | 72 |
| 18 | -3008 | 82 | -120 | 146 | 3008 | 210 | 120 |
| 19 | -2880 | 83 | -104 | 147 | 2880 | 211 | 104 |
| 20 | -2240 | 84 | -24 | 148 | 2240 | 212 | 24 |
| 21 | -2112 | 85 | -8 | 149 | 2112 | 213 | 8 |
| 22 | -2496 | 86 | -56 | 150 | 2496 | 214 | 56 |
| 23 | -2368 | 87 | -40 | 151 | 2368 | 215 | 40 |
| 24 | -3776 | 88 | -216 | 152 | 3776 | 216 | 216 |
| 25 | -3648 | 89 | -200 | 153 | 3648 | 217 | 200 |
| 26 | -4032 | 90 | -248 | 154 | 4032 | 218 | 248 |
| 27 | -3904 | 91 | -232 | 155 | 3904 | 219 | 232 |
| 28 | -3264 | 92 | -152 | 156 | 3264 | 220 | 152 |
| 29 | -3136 | 93 | -136 | 157 | 3136 | 221 | 136 |
| 30 | -3520 | 94 | -184 | 158 | 3520 | 222 | 184 |
| 31 | -3392 | 95 | -168 | 159 | 3392 | 223 | 168 |
| 32 | -22016 | 96 | -1376 | 160 | 22016 | 224 | 1376 |
| 33 | -20992 | 97 | -1312 | 161 | 20992 | 225 | 1312 |
| 34 | -24064 | 98 | -1504 | 162 | 24064 | 226 | 1504 |
| 35 | -23040 | 99 | -1440 | 163 | 23040 | 227 | 1440 |
| 36 | -17920 | 100 | -1120 | 164 | 17920 | 228 | 1120 |
| 37 | -16896 | 101 | -1056 | 165 | 16896 | 229 | 1056 |
| 38 | -19968 | 102 | -1248 | 166 | 19968 | 230 | 1248 |
| 39 | -18944 | 103 | -1184 | 167 | 18944 | 231 | 1184 |
| 40 | -30208 | 104 | -1888 | 168 | 30208 | 232 | 1888 |
| 41 | -29184 | 105 | -1824 | 169 | 29184 | 233 | 1824 |
| 42 | -32256 | 106 | -2016 | 170 | 32256 | 234 | 2016 |
| 43 | -31232 | 107 | -1952 | 171 | 31232 | 235 | 1952 |
| 44 | -26112 | 108 | -1632 | 172 | 26112 | 236 | 1632 |
| 45 | -25088 | 109 | -1568 | 173 | 25088 | 237 | 1568 |
| 46 | -28160 | 110 | -1760 | 174 | 28160 | 238 | 1760 |
| 47 | -27136 | 111 | -1696 | 175 | 27136 | 239 | 1696 |
| 48 | -11008 | 112 | -688 | 176 | 11008 | 240 | 688 |
| 49 | -10496 | 113 | -656 | 177 | 10496 | 241 | 656 |
| 50 | -12032 | 114 | -752 | 178 | 12032 | 242 | 752 |
| 51 | -11520 | 115 | -720 | 179 | 11520 | 243 | 720 |
| 52 | -8960 | 116 | -560 | 180 | 8960 | 244 | 560 |
| 53 | -8448 | 117 | -528 | 181 | 8448 | 245 | 528 |
| 54 | -9984 | 118 | -624 | 182 | 9984 | 246 | 624 |
| 55 | -9472 | 119 | -592 | 183 | 9472 | 247 | 592 |
| 56 | -15104 | 120 | -944 | 184 | 15104 | 248 | 944 |
| 57 | -14592 | 121 | -912 | 185 | 14592 | 249 | 912 |
| 58 | -16128 | 122 | -1008 | 186 | 16128 | 250 | 1008 |
| 59 | -15616 | 123 | -976 | 187 | 15616 | 251 | 976 |
| 60 | -13056 | 124 | -816 | 188 | 13056 | 252 | 816 |
| 61 | -12544 | 125 | -784 | 189 | 12544 | 253 | 784 |
| 62 | -14080 | 126 | -880 | 190 | 14080 | 254 | 880 |
| 63 | -13568 | 127 | -848 | 191 | 13568 | 255 | 848 |
Claims (10)
1. a kind of method for carrying out audio mixing to MCVF multichannel voice frequency based on programmable graphics processor, including：
The N roads voice data to be played in a period of time is stored to input texture buffer, the N roads voice data of input is tied up
It is set to the data texturing of programmable graphics processor；
Programmable graphics processor is pre-processed to each road voice data simultaneously, according to per audio all the way in preprocessing process
The coded system of data, sample rate or volume, the road voice data is decoded accordingly, in resampling or decay at least
One operation；Wherein, programmable graphics processor is carried out according to the coded system per voice data all the way to the road voice data
Corresponding decoding operate, corresponding re-sampling operations are carried out according to per the sample rate of voice data all the way to the road voice data,
Corresponding attenuation operations are carried out to the road voice data according to per the volume of voice data all the way；
Pretreated each road voice data is superposed to audio signal all the way by programmable graphics processor, by the road audio signal
Store to output texture buffer, and export to playback equipment.
2. method according to claim 1, it is characterised in that the programmable graphics processor is simultaneously to each road audio number
Also include according to before being pre-processed：
Seat of the sample value of the i-th tunnel input audio needed for calculating n-th output audio sample point in texture buffer is input into
Mark (x, y),
X=i
Y=n × ki/N
Wherein, kiThe number of elements of audio is input into for the i-th tunnel, i ∈ N, x are integer, and y is integer or decimal.
3. method according to claim 2, it is characterised in that the programmable graphics processor is according to per audio number all the way
According to sample rate, carrying out re-sampling operations to the road voice data includes：
Obtained from input texture buffer and be located at coordinateWithSample value at 2 points is respectively s1And s2,
Expression is rounded downwards,Expression rounds up；
If s1And s2Without coding, according to formulaInterpolation arithmetic is carried out to obtain (x, y)
Sample value s；
If s1And s2It is encoded, according to formulaInterpolation arithmetic is carried out to obtain (x, y)
Sample value s, s '1It is s1Decoded sample value, s'2It is s2Decoded sample value.
4. method according to claim 3, it is characterised in that if s1And s2It is encoded, carry out interpolation arithmetic it
Before, the programmable graphics processor carries out decoding behaviour according to the coded system per voice data all the way to the road voice data
Make, specifically include：
A rules or/and μ rule decodings look-up table array are bound the data texturing of programmable graphics processor；
Calculate audio sample values s1In the coordinate (u of a rules or μ rule decoding look-up table array correspondences texture buffer1, v1), u1=1,
v1=s1, according to coordinate (u1, v1) decode and obtain decoded sample value s by a rules or μ rule decoding look-up tables1'；
Calculate audio sample values s2In the coordinate (u of a rules or μ rule decoding look-up table array correspondences texture buffer2, v2), u2=1,
v2=s2, according to coordinate (u2, v2) decode and obtain decoded sample value s' by a rules or μ rule decoding look-up tables2。
5. method according to claim 3, it is characterised in that the programmable graphics processor is according to per audio number all the way
According to volume, carrying out decay to the road voice data includes：
The sample value s of (x, y) is multiplied by the attenuation factor of the i-th tunnel audioiAudio sample values after being decayed, and by after decay
Audio sample values be superimposed to output sample.
6. a kind of programmable graphics processor that audio mixing is carried out to MCVF multichannel voice frequency, including：
Data texturing binding module, for the N roads voice data to be played in a period of time to be stored to input texture buffer,
The N roads voice data of input is bound into the data texturing of programmable graphics processor；
Pretreatment module, for being pre-processed to each road voice data simultaneously, according to per audio all the way in preprocessing process
The coded system of data, sample rate or volume, the road voice data is decoded accordingly, in resampling or decay at least
One operation；Wherein, programmable graphics processor is carried out according to the coded system per voice data all the way to the road voice data
Corresponding decoding operate, corresponding re-sampling operations are carried out according to per the sample rate of voice data all the way to the road voice data,
Corresponding attenuation operations are carried out to the road voice data according to per the volume of voice data all the way；
Superposition output module, for pretreated each road voice data to be superposed into audio signal all the way, by the road audio letter
Number store to output texture buffer, and export to playback equipment.
7. programmable graphics processor according to claim 6, it is characterised in that the programmable graphics processor is also wrapped
Include：
Coordinate calculation module, for before being pre-processed to each road voice data, calculating to be exported for n-th needed for audio sample point
The i-th tunnel input audio sample value be input into texture buffer in coordinate (x, y),
X=i
Y=n × ki/N
Wherein, kiThe number of elements of audio is input into for the i-th tunnel, i ∈ N, x are integer, and y is integer or decimal.
8. programmable graphics processor according to claim 6, it is characterised in that the pretreatment module includes resampling
Unit, is used for
Obtained from input texture buffer and be located at coordinateWithSample value at 2 points is respectively s1And s2,
Expression is rounded downwards,Expression rounds up；
If s1And s2Without coding, according to formulaInterpolation arithmetic is carried out to obtain (x, y)
Sample value s；
If s1And s2It is encoded, according to formulaInterpolation arithmetic is carried out to obtain (x, y)
Sample value s, s '1It is s1Decoded sample value, s'2It is s2Decoded sample value.
9. programmable graphics processor according to claim 8, it is characterised in that the pretreatment module includes that decoding is single
Unit, if for s1And s2It is encoded, before interpolation arithmetic is carried out, a rules or/and μ rule decodings look-up table array are bound
The data texturing of programmable graphics processor；
Calculate audio sample values s1In the coordinate (u of a rules or μ rule decoding look-up table array correspondences texture buffer1, v1), u1=1,
v1=s1, according to coordinate (u1, v1) decode and obtain decoded sample value s ' by a rules or μ rule decoding look-up tables1；
Calculate audio sample values s2In the coordinate (u of a rules or μ rule decoding look-up table array correspondences texture buffer2, v2), u2=1,
v2=s2, according to coordinate (u2, v2) decode and obtain decoded sample value s' by a rules or μ rule decoding look-up tables2。
10. programmable graphics processor according to claim 8, it is characterised in that the pretreatment module includes decay
Unit, the attenuation factor for the sample value s of (x, y) to be multiplied by the i-th tunnel audioiAudio sample values after being decayed, and will
Audio sample values after decay are superimposed to output sample.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310170251.XA CN104143334B (en) | 2013-05-10 | 2013-05-10 | Programmable graphics processor and its method that audio mixing is carried out to MCVF multichannel voice frequency |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310170251.XA CN104143334B (en) | 2013-05-10 | 2013-05-10 | Programmable graphics processor and its method that audio mixing is carried out to MCVF multichannel voice frequency |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104143334A CN104143334A (en) | 2014-11-12 |
| CN104143334B true CN104143334B (en) | 2017-06-16 |
Family
ID=51852492
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310170251.XA Active CN104143334B (en) | 2013-05-10 | 2013-05-10 | Programmable graphics processor and its method that audio mixing is carried out to MCVF multichannel voice frequency |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104143334B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105976844A (en) * | 2016-04-29 | 2016-09-28 | 高翔 | Audio file generation method and device |
| CN109435837A (en) * | 2018-12-20 | 2019-03-08 | 云南玉溪汇龙科技有限公司 | A kind of engine of electric vehicle acoustic simulation synthesizer and method |
| CN110689876B (en) * | 2019-10-14 | 2022-04-12 | 腾讯科技（深圳）有限公司 | Voice recognition method and device, electronic equipment and storage medium |
| CN115312072B (en) * | 2022-08-04 | 2024-10-01 | 世邦通信股份有限公司 | Audio route management method, system and electronic equipment |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101123723A (en) * | 2006-08-11 | 2008-02-13 | 北京大学 | Digital video decoding method based on image processor |
| EP2184869A1 (en) * | 2008-11-06 | 2010-05-12 | Studer Professional Audio GmbH | Method and device for processing audio signals |
| CN102428514A (en) * | 2010-02-18 | 2012-04-25 | 杜比实验室特许公司 | Audio Decoder And Decoding Method Using Efficient Downmixing |
| CN102568481A (en) * | 2010-12-21 | 2012-07-11 | 富士通株式会社 | Method for implementing analysis quadrature mirror filter (AQMF) processing and method for implementing synthesis quadrature mirror filter (SQMF) processing |
| CN102638658A (en) * | 2012-03-01 | 2012-08-15 | 盛乐信息技术（上海）有限公司 | Method and system for editing audio-video |
| EP2544181A2 (en) * | 2011-07-07 | 2013-01-09 | Dolby Laboratories Licensing Corporation | Method and system for split client-server reverberation processing |
| CN102932645A (en) * | 2012-11-29 | 2013-02-13 | 济南大学 | Circuit structure integrating graphic processor and video codec |
| CN102968995A (en) * | 2012-11-16 | 2013-03-13 | 新奥特（北京）视频技术有限公司 | Sound mixing method and device of audio signal |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004208083A (en) * | 2002-12-25 | 2004-07-22 | Meiwa Yo | Method and device for producing contents of tv advertisement and inserting advertisement to tv program |
| US9942593B2 (en) * | 2011-02-10 | 2018-04-10 | Intel Corporation | Producing decoded audio at graphics engine of host processing platform |
-
2013
- 2013-05-10 CN CN201310170251.XA patent/CN104143334B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101123723A (en) * | 2006-08-11 | 2008-02-13 | 北京大学 | Digital video decoding method based on image processor |
| EP2184869A1 (en) * | 2008-11-06 | 2010-05-12 | Studer Professional Audio GmbH | Method and device for processing audio signals |
| CN102428514A (en) * | 2010-02-18 | 2012-04-25 | 杜比实验室特许公司 | Audio Decoder And Decoding Method Using Efficient Downmixing |
| CN102568481A (en) * | 2010-12-21 | 2012-07-11 | 富士通株式会社 | Method for implementing analysis quadrature mirror filter (AQMF) processing and method for implementing synthesis quadrature mirror filter (SQMF) processing |
| EP2544181A2 (en) * | 2011-07-07 | 2013-01-09 | Dolby Laboratories Licensing Corporation | Method and system for split client-server reverberation processing |
| CN102638658A (en) * | 2012-03-01 | 2012-08-15 | 盛乐信息技术（上海）有限公司 | Method and system for editing audio-video |
| CN102968995A (en) * | 2012-11-16 | 2013-03-13 | 新奥特（北京）视频技术有限公司 | Sound mixing method and device of audio signal |
| CN102932645A (en) * | 2012-11-29 | 2013-02-13 | 济南大学 | Circuit structure integrating graphic processor and video codec |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104143334A (en) | 2014-11-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106328126B (en) | Far field voice recognition processing method and device | |
| KR101771533B1 (en) | Method for Processing an Audio Signal, Signal Processing Unit, Binaural Renderer, Audio Encoder and Audio Decoder | |
| EP1921604B1 (en) | Environmental effects generator for digital audio signals | |
| KR101838623B1 (en) | Method for Processing an Audio Signal in accordance with a Room Impulse Response, Signal Processing Unit, Audio Encoder, Audio Decoder, and Binaural Renderer | |
| CN104143334B (en) | Programmable graphics processor and its method that audio mixing is carried out to MCVF multichannel voice frequency | |
| CN104904239B (en) | binaural audio processing | |
| EP1757165B1 (en) | Coding reverberant sound signals | |
| CN101379553B (en) | Apparatus and method for encoding/decoding signal | |
| CN110136744A (en) | A kind of audio-frequency fingerprint generation method, equipment and storage medium | |
| CN105981411A (en) | Multiplet-based matrix mixing for high-channel count multichannel audio | |
| CN104428834A (en) | Systems, methods, apparatus, and computer-readable media for three-dimensional audio coding using basis function coefficients | |
| TW569550B (en) | Method of inverse-modified discrete cosine transform and overlap-add for MPEG layer 3 voice signal decoding and apparatus thereof | |
| Tsingos | Precomputing geometry-based reverberation effects for games | |
| KR100760976B1 (en) | Computing circuits and method for running an mpeg-2 aac or mpeg-4 aac audio decoding algorithm on programmable processors | |
| CN117501362A (en) | Audio rendering system, method and electronic equipment | |
| CN102760437A (en) | Audio decoding device of control conversion of real-time audio track | |
| CN109087653B (en) | Method and apparatus for applying dynamic range compression to high order ambisonics signals | |
| US11705144B2 (en) | Methods and systems for encoding frequency-domain data | |
| KR20110033942A (en) | System, method or apparatus for combining multiple streams of media data | |
| CN102576531B (en) | Method and apparatus for processing multi-channel audio signals | |
| CN110534110A (en) | Robot and method, device and circuit for improving voice interaction recognition rate of robot | |
| CN117546236A (en) | Audio rendering system, method and electronic equipment | |
| CN109243472A (en) | A kind of audio-frequency processing method and audio processing system | |
| Du et al. | An implementation of filterbank for mpeg-2 aac on fpga | |
| CN202677855U (en) | Voice acquisition playback processing platform |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |